Combining Ability Studies for Development of New Hybrids in Rice over Environments

Combining ability study on grain yield and its components from line x tester analysis over the locations of five well adapted CMS lines and twenty three testers of different eco-geographic origin revealed higher SCA variance than GCA variance for all the traits indicating the prevalence of non-additive gene action. The lines APMS 6A, PUSA 5A and CRMS 32A and testers 1096, 1005, IBL-57 and SC5 9-3 were the good general combiners for yield and its majority of the traits. IBL-57 was the only good general combiner among the male parents for earliness, dwarfness and grain yield per plant. The hybrids APMS 6A x SC5 9-3, APMS 6A x 1005 and APMS 6A x GQ 25 were identified as potential one for yield and desired traits based on sca effects.


Introduction
The development and use of hybrid rice varieties on commercial scale utilizing CMS -fertility restoration system has proved to be one of the mile stones in the history of rice improvement.The hybrid rice technology now in operation, aims at yield increment through higher exploitable heterosis levels.With increasing interest in exploitation of heterosis in rice, there is an urgent need to subject various CMS lines and restorer lines for combining ability tests.The knowledge of combining ability is useful to assess nicking ability in self pollinated crops and at the same time elucidate the nature and magnitude of gene actions involved.It provides to the breeders an insight in to nature and relative magnitude of fixable and non-fixable genetic variances.Therefore, the present investigation was carried out to estimate combining ability effects for yield and its components involving CMS lines and restorer lines in rice.

Material and Methods
The material for the present study comprised 115 F 1 s of rice generated involving five CMS lines (viz., IR 58025A, IR 79156A, APMS 6A, PUSA 5A and CRMS 32A) and 23 diverse elite restorer lines as testers through Line x Tester design during rabi, 2006-07.The resultant 115 F 1 s and 28 parents were grown in randomized complete block design with two replications during kharif, 2007 at three different locations viz., Directorate of Rice Research, Hyderabad (Southern Telangana Agro-climatic zone), Regional Agricultural Research Station, Warangal (Central Telangana Agro-climatic zone) and Regional Agricultural Research Station, Jagtial (Northern Telangana Agro-climatic zone.Thirty days old seedlings were transplanted with one seedling per hill adopting 20 x 15 cm spacing and each entry was planted in two rows of 1.8 m length.All the recommended agronomic practices were followed.In each entry, five plants were selected randomly from each replication and biometrical observations were recorded for plant height, flag leaf length, flag leaf width, productive tillers per plant, panicle length, panicle weight, filled grains per panicle, spikelet fertility %, 1000 grain weight, grain yield per plant and productivity per day.Days to 50% flowering was recorded on plot basis.The mean data over the three locations were analyzed for combining ability following the standard method of Kempthorne (1957).

Results and Discussion
The pooled analysis of variance for combining ability over three locations revealed that the variance due to parents, parents vs. crosses and crosses were highly significant (Table 1) for all the characters studied.The variance due to lines was significant for all the characters except days to fifty per cent flowering, plant height and 1000 grain weight, whereas for testers, plant height, flag leaf length, productive tillers per plant, panicle length, filled grains per panicle and 1000 grain weight were found significant.Interaction effects of (Parents vs. crosses) x locations, parents x locations and crosses x locations were significant for all the characters, except flag leaf width in case of (parents vs. crosses) x locations interaction.Further partitioning of crosses x locations indicated that the interaction of lines x locations showed significant differences for productive tillers per plant, flag leaf length and flag leaf width, while testers x locations was significant only for filled grains per panicle.The interaction due to lines x testers and lines x testers x locations were significant for all the traits indicating that combining ability contributed heavily in the expression of these traits and provided importance of dominance or non additive variances for all the traits.The comparative estimates of variances due to GCA and SCA revealed the importance of SCA variance.The SCA variances were higher than GCA variances for all the traits suggesting the significant role of non additive gene action Predominance of non additive gene action for grain yield and its components was also reported by other workers (Satyanarayana et al.(2000), Rita and Motiramani (2005) ; Singh et al. (2005) ; Venkatesan et al. (2007) and Dalvi and Patel (2009) ) The estimates of gca effects (Table 2) showed that parents with high gca effects differed for various traits.Among the lines APMS 6A had favourable genes for grain yield and other traits including flag leaf length, flag leaf width, productive tiller per plant, panicle weight, filled grains per panicle, spikelet fertility % and productivity per day.The line PUSA 5A was the best combiner for grain yield, dwarfness, 1000 seed weight and productivity per day.However CRMS 32A contributed positive alleles for grain yield, dwarfness, panicle length, panicle weight, filled grains per panicle and productivity per day.Among the testers, 1096 possessed the desirable genes for grain yield, productive tillers per plant, panicle length, panicle weight, filled grains per panicle, and productivity per day, whereas 1005 for grain yield, dwarfness, flag leaf length, flag leaf width, panicle weight, filled grains per panicle and productivity per day.Some other good general combiners that had also contributed positive genes for various characters were; IBL-57 for grain yield, earliness of flowering, dwarfness, flag leaf length, productive tillers per plant, filled grains per panicle and productivity per day; SC5 9-3 for grain yield, flag leaf length, flag leaf width, productive tillers per plant, filled grains per panicle and spikelet fertility % and 124 for grain yield, panicle length, panicle weight, filled grains per panicle, spikelet fertility % and productivity per day.
Crosses with desirable sca effects for various traits along with mean performance and gca effects of parents involved in the crosses are listed in the table 3. The crosses APMS 6A X SC5 9-3, APMS 6A X 1005 and APMS 6A X GQ 25, PUSA 5A X IR 55 and PUSA 5A X IR 43 expressed significant sca effects as well as high per se performance for grain yield per plant and few other traits.However, CRMS 32A X IR 43 and PUSA 5A X IR 60 for plant height (dwarfness), IR 79156A X GQ-120 for flag leaf length, APMS 6A X 118 for flag leaf width, APMS 6A X SC 5 9-3 and APMS 6A X 118 for panicle weight, APMS 6A X 1005 for filled grains per panicle, APMS 6A X SC5 9-3 and APMS 6A X 1005 for grain yield per plant, PUSA 5A X KMR 3 for productivity per day had high mean performance and highly significant mean values.These desirable cross combinations involved high x high type of general combiners.Kalitha and Upadhaya (2000), Shivani et al. (2009) and Salgotra et al. (2009) also reported about interaction between positive and positive alleles in crosses involving high x high combiners which can be fixed in subsequent generations if no repulsion phase linkages are involved.
The desirable performance of combinations like high x low may be ascribed to the interaction between dominant alleles from good combiners and recessive alleles from poor combiners (Dubey, 1975).Such combinations were observed in the hybrids ; IR 58025AX BR 827-35 and IR 58025A X EPLT 109 for days to 50% flowering (earliness), IR 79156A X GQ 25 and IR 79156A X EPLT 109 for plant height (dwarfness), CRMS 32A X SC5 9-3, CRMS 32A X IBL-57 and IR 79156A x 1096 for flag leaf length, APMS 6A X 619-2 for flag leaf width, CRMS 32A X IR 43 for productive tillers per plant, IR 58025A XGQ 37-1, APMS 6A X GQ 120, IR 79156A X GQ 70 and IR 58025A X 517 for panicle length, IR 79156A X SG 27-77 for filled grains per panicle, APMS 6A X 611-1 for spikelet fertility %, PUSA 5A X 517, APMS 6A X IR 60 and APMS 6A X EPLT 109 for 1000 grain weight, APMS 6A X GQ 25, PUSA 5A X IR 43 and PUSA 5A XIR 55 for grain yield per plant and APMS 6A X GQ 25, PUSA 5A X IR 43, PUSA 5A X IR 55 and IR 79156A X IBL 57 for productivity per day.Peng and Virmani (1990) also reported the possibility of interaction between positive alleles from good combiner and negative alleles from poor combiner in high x low cross combination and suggested for the exploitation of F 1 generation, as their high yielding potential would be unfixable in succeeding generation.
Combining ability analysis revealed that 1096, 1005, IBL-57 and SC5 9-3 were the good general combiners for yield and its majority of the trait among the male lines, whereas APMS 6A, PUSA 5A and CRMS 32A were among female lines.APMS 6A x SC5 9-3, APMS 6A x 1005 and APMS 6A x GQ 25 were identified as most promising crosses for yield based on sca effects, better per se and both or one of the parents with high gca for yield per plant also, could be exploited profitably for yield in rice.

Table 1a .
Pooled analysis of variance for combining ability (L X T) for yield and yield components in rice

Table 1b .
Pooled analysis of variance for combining ability (L X T) for yield and yield components in rice

Table 3 .
Top five crosses with high sca effects, per se performance and gca effects of parents for grain yield and